Separation in Other Coordinate Systems
You'll learn how to apply separation of variables to PDEs in polar, cylindrical, and spherical coordinates when solving problems on circular or radial domains.
Why Change Coordinates?
Up to now, you've used separation of variables in Cartesian coordinates (x, y, z). But many physical problems naturally fit circular or spherical geometries—like heat diffusion in a disk, vibrations of a drum, or electrostatic potential around a sphere. In these cases, using the problem's natural symmetry makes separation cleaner and the boundary conditions simpler.
The Core Idea
When you transform a PDE into polar (r, θ), cylindrical (r, θ, z), or spherical (r, θ, φ) coordinates, the Laplacian and other operators change form. You then assume a product solution that respects the coordinate structure:
- Polar: u(r, θ) = R(r)Θ(θ)
- Cylindrical: u(r, θ, z) = R(r)Θ(θ)Z(z)
- Spherical: u(r, θ, φ) = R(r)Θ(θ)Φ(φ)
Substituting this product form into the PDE (say, Laplace's equation in polar coordinates) and dividing by the product separates the equation into independent ODEs—just like before, but now in the new coordinate variables.
What Changes?
The resulting eigenvalue problems differ from Cartesian cases:
- In polar/cylindrical, you often get Bessel's equation for the radial part R(r)
- Angular parts yield periodic conditions (e.g., Θ(0) = Θ(2π))
- In spherical, you encounter Legendre's equation for the polar angle θ
These specialized ODEs have their own families of solutions (Bessel functions, Legendre polynomials), which replace the sines and cosines you used in rectangular domains.
Key Takeaway: Separation of variables extends naturally to polar, cylindrical, and spherical coordinates by assuming product solutions matched to the geometry; this transforms the PDE into ODEs whose solutions involve special functions tailored to radial and angular symmetry.